INDEX

Explanations

words related to a specific name, "Rhonda."

oai_token-act-pair · gpt-3.5-turbo

the name "Rhonda."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

cry

-0.66

 Hebrew

-0.64

Eps

-0.64

 servers

-0.63

 probability

-0.63

 Prof

-0.60

 draw

-0.60

 spread

-0.60

 movement

-0.60

 materials

-0.60

POSITIVE LOGITS

onda

4.87

ond

1.36

onde

1.28

unda

1.13

ondo

1.07

ayne

1.07

iott

1.01

 Honda

1.00

issan

1.00

orthy

0.98

Activations Density 0.020%

words related to a specific name, "Rhonda."

the name "Rhonda."

No Comments

No Known Activations